Hi Urs, I do not know when will we end up meeting (actually, Daniele has just posted the web page with the material to take the decisions, so I am afraid we will have to postpone since you did not have enough time to digest it) I would like to comment on your web page anyhow. 1) I apologize for not having cleaned up, but the comparison table at the bottom of my page is obsolete and it does not refer to any of the finally proposed selections. (while the page is uptodate on the reco vs recoil study). I sort of missed the point on why you discuss the comparisons at the bottom of my page if you then redid all the plots yourself. Anyhow Daniele was meant to circulate the final set of plots but the afs glitch of yesterday delayed him. 2) let's then discuss your plots (which should be the same as daniele's, I hope we do not have to start a saga about differences between the two sets of allegedly identical plots). * mxhadfit : we all agree that on depleted things are definitely better. On the enriched we should not care about the agreement otherwise we will measure the MC number by definition... * mm2 : here the situation is annoying. I am happy that we have found a way to perturbate the data-MC agreement. The improvement in the depleted sample is out of discussion, and maybe this should close all discussions. On the other side in the enriched sample the negative side flips over: data were above and go significantly below. Should we care about this? Should we do the systematics varying between the two selections? Should we adopt phot#5 which is intermediate? Maybe we should just do the MM2 scan of the measurement in the three cases and decide upon what we find. * Nneu : the annoying part here is that before the cuts the shapes are reasonably similar between data and MC while after the cuts the shapes do not talk to each other in the enriched sample. This is not improved by the new selections, it really looks like related to physics. * hadron spectra Your conclusion that there is still a low mom problem comes from http://www.slac.stanford.edu/~ursl/talks/092002/anaQA-a02/bmc-a-a5070.eps I guess, which is bit supported by Daniele's corresponding plot in http://www.slac.stanford.edu/BFROOT/www/Physics/Analysis/AWG/InclusiveSL/islrecoil/datagene0909_80fb-1_hist/a1610.eps In summary, IMVHO: - if we do not have other striking ideas about studies to be done, we have to go ahead, redo optimization and fitting in all subsamples. The only decision to be taken is whether to adopt phot5 or phot6 as default. If we adopt the criteria that we should look only at depleted for decisions, phot6 is definitely better. - I agree with the statement that we have not found the solution to the problems of the whole world. Since I like positive statements, I think that we have found a mechanism (photon selection) that has a large impact on the distribution that showed the most disturbing disagreement. I call this "better understanding". - since I do not like negative attitudes, my 0.02$ are that in July we missed the goal by little and by bad luck. If this statement is true, with the new selection, improved understanding and enhanced statistics the tests that were not satisfactory should not fail any more. I therefore believe we have to perform all the studies in july asap. If everything goes smooth (which I think is not unlikely) go ahead and get the number out asap. If there are still problems, play with the handles we now have to understand the situation ... ciao ric On Thu, 19 Sep 2002, Urs Langenegger wrote: > > Hoi, > > I am open to any time without meeting overlap, the following is just > for completeness: > > > Call 1-510-647-3480, press 1, enter 220824# and follow the instructions to attend. > > Time: 09-20-2002 08:30 AM (US/Pacific) > > > My preliminary contributions are being assembled in > > http://www.slac.stanford.edu/~ursl/talks/092002 > > Cheers, > --U. >